Performance Evaluation of

FR welded connections to Inclined Column

Yong-Wan Kim
Research Engineer, Chang Minwoo Structural Consultants, Inc., Seoul, Korea
(ywkim@minwoo21.com)

Jin Sun Park

Research Engineer, Chang Minwoo Structural Consultants, Inc., Seoul, Korea

Yongjei Lee, Ph. D., P.E.

Senior Associate, Chang Minwoo Structural Consultants, Inc., Seoul, Korea

Taejin Kim, Ph.D., P.E.

Executive Director, Chang Minwoo Structural Consultants, Inc., Seoul, Korea

SUMMARY
As the complex-shaped buildings become a popular trend, more researches are needed to embody those twisted
and tilted shapes into real structures. The studies on the inclined column and beam connection that appears
frequently in the complex-shaped structures are not sufficient in comparison to those on the conventional
structures so that the structural safety and the behavior are not clear yet. In this study, with a help of finite
element analysis, the fracture potential and local buckling which can occur around the connection are examined
and the results will be applied to evaluate the connection capacity. Also, the pre-analysis results will be used in
the future experimental studies.
Keywords: Inclined column, welded connection, finite element analysis, fracture, local buckling

1. INTRODUCTION
Nowadays, it is a new trend of architecture and structural engineering that many tall buildings on all
over the world have a complex shape such as twisted, tapered and tilted. Those buildings are regarded
as landmarks of a nation and symbols of technological achievement. The complex-shaped tall building
disassembles the orthogonality between beams and columns. Especially, such type of structural system
can be found frequently in buildings having inclined columns. The inclined columns can transmit the
gravity and lateral load simultaneously, and are necessary to display the architectural complexity.
Nonetheless, the studies for the connection of inclined columns have not been conducted as much as
those for the normal beam-column connections (Kim et al., 2011), so it is worth examining the
behavior of inclined column-beam connection by using rigorous computational analysis and
experiments.
In this study, a moment resisting frame which has been used for SAC project conducted in the United
States (Gupta et al., 1999) was used as an model building frame. The model building was modified to
have inclined columns at one side of its ground floor, and was analyzed with gravity loads to check the
flow of forces at the inclined column-beam connections. After the global analysis of the example
frame, which revealed that the inclined column gives an extra axial force to the beam, the finite
element analysis (FEA) was conducted to examine the behavior of inclined column connection more
precisely. For the FEA, three types of column-beam connection substructure were developed so that
the two different inclined directions of column could be compared to the normal column-beam
connection.
Finally, this study provides the prediction on the behavior of welded inclined column connection
focused on the strain distribution and local buckling so that the fracture potential at the connection and
strength degradation can be assessed.

2. MODEL BUILDING
The structural plan of the model building is shown in Fig. 1(a). According to SAC project, it was
assumed that the building is located in LA, having 20 stories and steel moment resisting frame system.
The girders and columns in exterior frames are connected by fully restrained welded connections. For
the modeling and elastic analysis of the model building, ETABS, a structural analysis software
developed by Computers & Structures, Inc., was used. During the 3D modeling, the columns in the
west side of the building on the 2nd floor, which are surrounded by a dashed box in Fig. 1(a), were
modified to be inclined (see Fig. 1(b)). The building was analyzed for two different angles of the
inclined columns, 1/10 and 1/5, which are equal to 5.7 and 11.3 from vertical.
Since LA is a region of high seismicity, in the SAC project the structural design of the model building
was governed by earthquake load, rather than wind load. Therefore, the model building was analyzed
with seismic load as well as gravity load based on both UBC 94 and ASCE 7-10, and each frame
member of the building was checked for demand/capacity ratio in terms of moments and forces. It is
notable that the model building has been developed to represent pre-Northridge structures designed
based on the UBC 94 which categorizes all moment resisting frames as the one identical system
having a response modification factor of 12. The summary of seismic design information for model
building is provided in Table 1.
Table 1. Seismic design information of the model building based on UBC 94
System
Steel Moment Resisting Frame
R factor (RW)

12

Soil Type

S2 (Stiff Soil)

Zone Factor (Z)

0.4 (Zone 2)

Importance Factor (I)

Period (TA)

2.3 sec

Design Base Shear Coefficient (C)

Seismic Weight

0.03
101722 kN

As a result of elastic analysis, it was found that there is a tensile axial force in the beam growing up
proportional to the inclined angle of column.

(a) Plan

(b) Elevation
Figure 1. Model building frames

3. FINITE ELEMENT MODELS

After the global analysis, a substructure model of the column-beam connection was developed for
FEA. For the substructure model, a typical inclined column-beam connection on the second floor was
chosen to be used. Table 2 shows material properties of frame members of the inclined column-beam
substructure.
Table 2. Material properties of columns and beam used in analysis model
Youngs
Yield
Tensile
Member
Cross section
Modulus
Strength
Strength
(MPa)
(MPa)
(MPa)
Column
H-417x458x30x50
205,000
345
448
Beam

H-900x300x16x28

205,000

248

Elongation
(%)

Poissons
ratio

21

0.3

23

0.3

414

The boundary and loading conditions for the analysis model of the beam-inclined column connection
were set as shown in Fig. 2, considering the conditions for experimental study which will be
conducted in the future. The analysis model includes the substructure of the model building having the
inclined column-beam connection, disassembled by the cut at the inflection point in each components.
The inflection points were restrained with a pin and a roller, and the beam was constrained against
their out-of-plane behavior. The vertical displacement at the end of the beam was controlled
monotonically by using a unit load and its proportionality factor.

Figure 2. Boundary and loading conditions

As expected in the global analysis, the horizontal force in the beam is developed from the axial force
in the inclined column as shown in Fig. 3. However, since the beam is constrained by the slab which is
expected to take most of the horizontal force, only a little region may be affected so it becomes a
challenge in analysis of the substructure. Therefore, to determine the region which the horizontal force
affects the beam only, an effective region was assumed. The effective region includes the beam web
and flanges from the column face to a half of the beam depth (see Fig. 4).

(a) BC-T

(b) BC-C

(c) BC-N

Figure 3. Nodal forces in inclined column connections

In addition, for more precise investigation of the effect of horizontal distribution of column axial force
to the beam, it was decided to apply axial stresses in the effective zone as an initial condition,
corresponding to the columns having different inclined angles, rather than applying an axial force
directly to the column.

(a) Effective region

(b) Initial stress assignment

Figure 4. Effective region of beam and an example of initial stress assignment

The finite element analysis (FEA) was performed in ABAQUS. For the FE substructure modeling, it
was assumed that Welded Unreinforced Flange-Welded Web (WUF-W) method was used for the
connection without weld access holes, and shell elements having four nodes for reduced integration
method (S4R) were assigned. Since the substructure model obtained from the model building has only
tensile force in the beam as shown in Fig. 3(a), another case shown in Fig. 3(b) was suggested to
consider the compressive force. Also, the performance of the beam-inclined column connections was
compared to the vertical column-beam connection (see Fig. 3(c)) with same conditions.

4. ANALYSIS RESULTS
Since it was anticipated that the axial stress in the beam distributed from the inclined column would
affect the beam-column behavior such as the potential of fracture at the column face and local
buckling in beam flange, the finite element analysis focused on stress-strain and force-deformation
relationship of substructure model. For the fracture potential, the plastic equivalent strain(PEEQ) at
the beam flange was checked, while the relationship between the vertical displacement and applied
load at the end of the beam was examined to check the flange local buckling(FLB). The PEEQ was
measured in the effective region of top and bottom flanges, and the FLB was checked by using
eigenvalue analysis and displacement controlled analysis with Riks method provided in ABAQUS
(ABAQUS, 2010).
4.1. Fracture Potential
El-Tawil et al.(1998) and Yoo(2006) has studied on the PEEQ of a steel component as an index of the
local failure, and proved that large PEEQ indicates the high potential of fracture. Referring to those
previous research achievements, this study also measured PEEQ to assess the fracture potential as well
as the yielding occurrence. The applied initial stress in the beam which acts as a distributed column
axial force will obviously give extra deformation of beam flanges and web. Consequently, the beam
will have larger plastic strain than vertical column-beam connection, which means that yielding will
occur earlier and fracture potential will be higher, when there is a same amount of beam deflection.
Therefore, as a criterion for the assessment of the effect of inclined column on the yielding and
fracture potential at the connection, PEEQ was checked in ABAQUS, with tip vertical displacement of
70mm which was thought to be a reasonable condition to see yielding at column face in every case.
Since it was predicted that the top and bottom flanges would go over the largest stress and strain when

the beam deflects under the gravity load, variation of the PEEQ was examined along the effective
region of the beam at the middle of the flange width as shown in Fig. 5.
As mentioned earlier, the forces in the beam were supposed to be applied according to the horizontal
distribution of axial forces in the inclined column. However, 25% and 50% of yield strength of the
beam were applied in order to observe the amplified results, instead of considering realistic or
practical inclined angle of the column which gives a force of less than 20% of beam yielding strength.
One can see the distribution of Von Mises stress around the beam flange and column face in Fig. 5(a)
and the deformed shape of substructure in Fig. 5(b), which were captured when yielding begins to
occur. The overall schedule for the FEA is listed in Table 3.

(a) Von Mises stress distribution

(b) Deformed shape at onset of yielding

Figure 5. Stress distribution and deformed shape of substructure model when yielding occurs
Table 3. Analysis models
Model name
BC-T-25
BC-T-50
BC-C-25
BC-C-50
BC-N

Direction of axial force

Tension
Compression

Amount of axial force (stress)

25% fy
50% fy
25% fy
50% fy

No initial stress

Not applicable

From Fig. 6, it can be noticed that, if the top flange of the beam carries tensile stress due to the flexure,
in case of that the horizontally distributed column axial force is a tensile force (BC-T), the plastic
deformation occurs more severely than that in the normal column-beam connection (BC-N) at the
same amount of beam deflection. It is also possible to point out from Fig. 6 that, as expected earlier,
the tensile initial stress causes larger potential of fracture since the PEEQ at the column face increases
in proportion to the inclined angle.

Figure 6. PEEQ distribution in the top flange of beam

On the other hand, as Fig. 7 shows, in case of that a compressive axial force acts in the beam (BC-C),
the top flange undergoes smaller plastic deformation than the normal one.

Figure 7. PEEQ distribution in the bottom flange of beam

Therefore, it can be concluded that, if the horizontal distribution of an inclined column acts in a beam
with flexure, the smaller or larger deflection under the same deflection is expected comparing to the
vertical column connection, depending on the inclined angle and direction of column. Although this
study used extreme cases for the analysis, the results still provide the prediction of a change in plastic
strain distribution which affects the local failure. Furthermore, the possible change in plastic hinge
rotation can be expected.
4.2. Local Buckling
Due to the fact that the axial stress acting in the beam will cancel or amplify the compressive stress
which grows during the flexural deformation depending on its direction, the onset of the FLB in the
beam can be delayed or advanced. Also, once the local buckling occurs, the column-beam connection
will undergo the strength degradation because of its capacity loss. Therefore, to verify and visualize
this effect, by using ABAQUS, the eigenvalue analysis was conducted considering the buckling mode
as shown in Fig. 8(a), then the displacement controlled analysis using Riks method, which is a
nonlinear analysis considering the buckling, was performed.
For the eigenvalue and Riks analysis, the upward displacement at the end of the beam was applied and
increased until the tip displacement reaches 500mm, which induces about 0.12 radians of end rotation
at the column face and thought to be enough to see the FLB in every case. The initial imperfection was
defined as 1/10 of the beam flange. Fig. 8(b) describes buckled shape of the beam, and Fig. 9 and Fig.
10 show the buckling analysis results of each model.

(a) Flange local buckling mode

(b) Buckled shape of beam

Figure 8. Flange local buckling mode and buckled shape of beam

To compare the occurrence of flange local buckling in beams connected to columns having different
angle, 25%, 50% and additionally 75% of yield strength of beam in tension and compression, which
also represent extraordinary amount of inclined angle of column, were applied. Fig. 9 demonstrates the
relationship between the upward displacement and the vertical force at the end of the beam when the
tensile stress is acting in the beam. In contrast, Fig. 10 shows the same plot from the beam having
compressive axial stress.
From Fig. 9, it can be found that, as the axial tensile stress in the beam increases, a plateau in the curve
appears at larger tip displacement, which means that the onset of buckling is delayed as the inclined
angle of column increases, while Fig. 10, a case of compressive stress in the beam, shows the reversed
result. Also, it shall be noted that, as the inclined angle of column increases, the initial stiffness of the
inclined column-beam substructures is reduced earlier because of the advanced yielding in the beam
flange due to the extra axial stress. Therefore, when a steel beam connected to the inclined column is
designed, the limit state for the FLB and the elastic stiffness should be considered carefully.

Figure 9. Load-displacement relationship of beam with tensile stress considering FLB

Figure 10. Load-displacement relationship of beam with compressive stress considering FLB

5. CONCLUSION
As described above, there exists a tensile or compressive stress in the beam connected to the inclined
column, increasing with respect to the inclined angle. This horizontally distributed axial force from the
inclined column affects the stress and strain distribution in the beam directly, so the resisting capacity
of beam against the flexural deformation changes.
In this study, the plastic strain distribution and local buckling in beam flange were examined to assess
the seismic performance of inclined column-beam connection. Although the amplified amount of force

was applied to the model, which means that the inclined angle of column is unrealistic, the results of
this study proved that the horizontal distribution of the axial force in the inclined column affects the
potential of fracture at the connection by causing early yielding, and that it also delays or facilitates
flange local buckling. Therefore, it is possible to conclude that the extra check for the inelastic
behavior, local failure and buckling shall be considered when the inclined column-beam connection is
designed.
Since this study deals only with the strain distribution and local buckling of beam flanges, it is still
necessary to check fatigue at the connection, because they are also crucial for the seismic performance
of structural members, and sensitive to the amount of axial stress around the welded connection joint.
Furthermore, a full scale experiment shall be performed to prove the prediction of behavior from finite
element analysis.
ACKNOWLEDGEMENT
This research was supported by a grant (Code# 09 R&D A01) from Cutting-edge Urban Development Program
funded by Ministry of Land, Transport and Maritime Affairs of Korean Government.
REFERENCES
Gupta, A. and Krawinkler, H.(1999). Seismic Demands for Performance Evaluation of Steel Moment Resisting
Frame Structures, The John A. Blume Earthquake Engineering Center, Report No. 132, Stanford University, CA
American Society of Civil Engineers(ASCE) (2010). Minimum Design Loads for Buildings and Other Structures,
ASCE 7-10, ASCE, Reston, VA
International Conference of Building Official(ICBO) (1994). Uniform Building Code 1994 Edition, Vol. 2,
ICBO, Whittier, CA
American Institute of Steel Construction(AISC) (2006). Steel Construction Manual, 13th Edition, AISC,
Chicago, IL
ABAQUS (2010), ABAQUS/Standard Users Manual version 6.11, ABAQUS, Inc.
Kim, Y. W., Kim. S. and Kim. J.(2011). Analytical Study on Structural Behavior of Beam-Column Connections
with Inclined Columns, Proceedings of 22nd Annual Academic Conferences, Korean Society of Steel
Construction, 2011, 141 (in Korean)
Kim, E. S., Cho, J. H., Kim, T. and Kim, J. H.(2010). Structural Performance Evaluation of Connection with
Inclined Column for Steel Structures Subjected to Vertical Loads, Proceedings of Annual Conferences,
Computational Structural Engineering Institute of Korea, 2011-04 (in Korean)
Lu, L. W., Ricles, J. M., Mao, C. and Fisher, J. W.(2000). Critical Issues in Achieving Ductile Behaviour of
Welded Moment Connections, Journal of Constructional Steel Research 55(2000), 325341
Ricles, J. M., Mao, C., Lu, L. W. and Fisher, J. W.(2002). Inelastic Cyclic Testing of Welded Unreinforced
Moment Connections, Journal of Structural Engineering, Vol. 128, No. 4, April 2002, 429-440,
Kim, T.(2003). Experimental and Analytical Performance Evaluation of Welded Steel Moment Connections to
Box or Deep W-shape Columns, Ph.D. thesis, Dept. of Civil and Environmental Engineering, University of
California at Berkeley
El-Tawil, S., Mikesell, T., Vidarsson, E. and Kunnath, S. K.(1998). Strength and Ductility of FR Welded-Bolted
Connections, SAC Joint Venture, Report No. SAC/BD-98/01, Sacramento, CA
Yoo, J.(2006). Analytical Investigation on the Seismic Performance of Special Concentrically Braced Frames,
Ph.D. thesis, Dept. of Civil and Environmental Engineering, University of Washington